Tree-structured neural networks: Spatiotemporal dynamics and optimal control.

How the network topology drives the response dynamic is a basic question that has not yet been fully answered in neural networks. Elucidating the internal relation between topological structures and dynamics is instrumental in our understanding of brain function. Recent studies have revealed that the ring structure and star structure have a great influence on the dynamical behavior of neural networks. In order to further explore the role of topological structures in the response dynamic, we construct a new tree structure that differs from the ring structure and star structure of traditional neural networks. Considering the diffusion effect, we propose a diffusion neural network model with binary tree structure and multiple delays. How to design control strategies to optimize brain function has also been an open question. Thus, we put forward a novel full-dimensional nonlinear state feedback control strategy to optimize relevant neurodynamics. Some conditions about the local stability and Hopf bifurcation are obtained, and it is proved that the Turing instability does not occur. Moreover, for the formation of the spatially homogeneous periodic solution, some diffusion conditions are also fused together. Finally, several numerical examples are carried out to illustrate the results' correctness. Meanwhile, some comparative experiments are rendered to reveal the effectiveness of the proposed control strategy.

Stability and Bifurcation Exploration of Delayed Neural Networks With Radial-Ring Configuration and Bidirectional Coupling.

For decades, studying the dynamic performances of artificial neural networks (ANNs) is widely considered to be a good way to gain a deeper insight into actual neural networks. However, most models of ANNs are focused on a finite number of neurons and a single topology. These studies are inconsistent with actual neural networks composed of thousands of neurons and sophisticated topologies. There is still a discrepancy between theory and practice. In this article, not only a novel construction of a class of delayed neural networks with radial-ring configuration and bidirectional coupling is proposed, but also an effective analytical approach to dynamic performances of large-scale neural networks with a cluster of topologies is developed. First, Coates' flow diagram is applied to acquire the characteristic equation of the system, which contains multiple exponential terms. Second, by means of the idea of the holistic element, the sum of the neuron synapse transmission delays is regarded as the bifurcation argument to investigate the stability of the zero equilibrium point and the beingness of Hopf bifurcation. Finally, multiple sets of computerized simulations are utilized to confirm the conclusions. The simulation results expound that the increase in transmission delay may cause a leading impact on the generation of Hopf bifurcation. Meanwhile, the number and the self-feedback coefficient of neurons are also playing significant roles in the appearance of periodic oscillations.

Machine learning: A non-invasive prediction method for gastric cancer based on a survey of lifestyle behaviors.

Gastric cancer remains an enormous threat to human health. It is extremely significant to make a clear diagnosis and timely treatment of gastrointestinal tumors. The traditional diagnosis method (endoscope, surgery, and pathological tissue extraction) of gastric cancer is usually invasive, expensive, and time-consuming. The machine learning method is fast and low-cost, which breaks through the limitations of the traditional methods as we can apply the machine learning method to diagnose gastric cancer. This work aims to construct a cheap, non-invasive, rapid, and high-precision gastric cancer diagnostic model using personal behavioral lifestyles and non-invasive characteristics. A retrospective study was implemented on 3,630 participants. The developed models (extreme gradient boosting, decision tree, random forest, and logistic regression) were evaluated by cross-validation and the generalization ability in our test set. We found that the model developed using fingerprints based on the extreme gradient boosting (XGBoost) algorithm produced better results compared with the other models. The overall accuracy of which test set was 85.7%, AUC was 89.6%, sensitivity 78.7%, specificity 76.9%, and positive predictive values 73.8%, verifying that the proposed model has significant medical value and good application prospects.

Application of machine learning algorithms in predicting HIV infection among men who have sex with men: Model development and validation.

Background: Continuously growing of HIV incidence among men who have sex with men (MSM), as well as the low rate of HIV testing of MSM in China, demonstrates a need for innovative strategies to improve the implementation of HIV prevention. The use of machine learning algorithms is an increasing tendency in disease diagnosis prediction. We aimed to develop and validate machine learning models in predicting HIV infection among MSM that can identify individuals at increased risk of HIV acquisition for transmission-reduction interventions. Methods: We extracted data from MSM sentinel surveillance in Zhejiang province from 2018 to 2020. Univariate logistic regression was used to select significant variables in 2018-2019 data (P < 0.05). After data processing and feature selection, we divided the model development data into two groups by stratified random sampling: training data (70%) and testing data (30%). The Synthetic Minority Oversampling Technique (SMOTE) was applied to solve the problem of unbalanced data. The evaluation metrics of model performance were comprised of accuracy, precision, recall, F-measure, and the area under the receiver operating characteristic curve (AUC). Then, we explored three commonly-used machine learning algorithms to compare with logistic regression (LR), including decision tree (DT), support vector machines (SVM), and random forest (RF). Finally, the four models were validated prospectively with 2020 data from Zhejiang province. Results: A total of 6,346 MSM were included in model development data, 372 of whom were diagnosed with HIV. In feature selection, 12 variables were selected as model predicting indicators. Compared with LR, the algorithms of DT, SVM, and RF improved the classification prediction performance in SMOTE-processed data, with the AUC of 0.778, 0.856, 0.887, and 0.942, respectively. RF was the best-performing algorithm (accuracy = 0.871, precision = 0.960, recall = 0.775, F-measure = 0.858, and AUC = 0.942). And the RF model still performed well on prospective validation (AUC = 0.846). Conclusion: Machine learning models are substantially better than conventional LR model and RF should be considered in prediction tools of HIV infection in Chinese MSM. Further studies are needed to optimize and promote these algorithms and evaluate their impact on HIV prevention of MSM.

Liquid chromatography-tandem mass spectrometric assay for TPN171 in human plasma.

A simple, rapid and accurate method for quantitative analysis of a highly selective phosphodiesterase-5 inhibitor (PDE5), TPN171 by high performance liquid chromatography and tandem mass spectrometry in human plasma was proposed and validated successfully using D3-TPN171 as internal standards (ISTD). An aliquot of 100 µL of plasma was mixed with internal standard and was precipitated with acetonitrile. Gradient elution was performed on a ACQUITY HSS T3 column (50 × 2.1 mm, 1.8 µm) coupled with a ACQUITY column in-line filter at 40℃, by 5 mM ammonium acetate in water containing 0.1 % formic acid and 0.1 % formic acid in acetonitrile as the mobile phase. The total analytical run time was 3.5 min. The analyte was monitored using multiple reaction monitoring (MRM) scan in positive polarity mode. The ion transition was m/z 442.2→113.2 and 445.2→116.2 for TPN171 and D3-TPN171 respectively. The method was validated for specificity, sensitivity, precision, accuracy, and other analytical parameters. The results found were satisfactory over the linear calibration range of 1-500 ng/mL. Within-day precisions ranged from 1.8 to 7.3 %, and between-day precisions from 2.3 to 4.9 %, accuracies were 95.5-99.8 %.The validated method was successfully applied to determine the plasma concentration after oral administration of 10 mg TPN171 in six healthy volunteers.

[Definition and function identification of nucleus export signal of BRD7].

OBJECTIVE: To localize and define the region of nucleus export signal (NES) on BRD7, and determine the role of this region in nucleus export of the external protein. METHODS: Based on an in vitro expressed model of green fluorescence protein (GFP), we performed DNA walking analysis to set BRD7 into several sections according to the structural characteristics of BRD7, investigated the effect of different sections of BRD7 on nucleus export of GFP, defined the region of nucleus export signal sequence of BRD7, and further ascertained the content of amino acids in BRD7 and potential localization of BRD7 NES by bioinformatics. RESULTS: B7C1 fragments ranged from aa219 to aa450 in BRD7 were found to target the external protein GFP into the cytoplasm detected by GFP direct fluorescence, which could be inhibited by NES inhibitor Leptomycin B (LMB). This region was rich in hydrophobic amino acid residues but no typical NES with characteristics of leucine-rich sequence by bioinformatics. CONCLUSION: The region from aa219 to aa450 is primarily defined as an atypical NES in BRD7.

Transcriptional regulation of BRD7 expression by Sp1 and c-Myc.

BACKGROUND: Bromodomain is an evolutionally conserved domain that is found in proteins strongly implicated in signal-dependent transcriptional regulation. Genetic alterations of bromodomain genes contributed to the development of many human cancers and other disorders. BRD7 is a recently identified bromodomain gene. It plays a critical role in cellular growth, cell cycle progression, and signal-dependent gene expression. Previous studies showed that BRD7 gene exhibited much higher-level of mRNA expression in normal nasopharyngeal epithelia than in nasopharyngeal carcinoma (NPC) biopsies and cell lines. However, little is known about its transcriptional regulation. In this study, we explored the transcriptional regulation of BRD7 gene. METHOD: Potential binding sites of transcription factors within the promoter region of BRD7 gene were predicted with MatInspector Professional http://genomatix.de/cgi-bin/matinspector_prof/mat_fam.pl. Mutation construct methods and luciferase assays were performed to define the minimal promoter of BRD7 gene. RT-PCR and western blot assays were used to detect the endogenous expression of transcription factor Sp1, c-Myc and E2F6 in all cell lines used in this study. Electrophoretic mobility shift assays (EMSA) and Chromatin immunoprecipitation (ChIP) were used to detect the direct transcription factors that are responsible for the promoter activity of BRD7 gene. DNA vector-based siRNA technology and cell transfection methods were employed to establish clone pools that stably expresses SiRNA against c-Myc expression in nasopharyngeal carcinoma 5-8F cells. Real-time PCR was used to detect mRNA expression of BRD7 gene in 5-8F/Si-c-Myc cells. RESULTS: We defined the minimal promoter of BRD7 gene in a 55-bp region (from -266 to -212bp), and identified that its promoter activity is inversely related to c-Myc expression. Sp1 binds to the Sp1/Myc-Max overlapping site of BRD7 minimal promoter, and slightly positively regulate its promoter activity. c-Myc binds to this Sp1/Myc-Max overlapping site as well, and negatively regulates the promoter activity and endogenous mRNA expression of BRD7 gene. Knock-down of c-Myc increases the promoter activity and mRNA level of BRD7 gene. The luciferase activity of the mutated promoter constructs showed that Sp1/Myc-Max overlapping site is a positive regulation element of BRD7 promoter. CONCLUSION: These studies provide for the first time the evidence that c-Myc is indeed a negative regulator of BRD7 gene. These findings will help to further understand and uncover the bio-functions of BRD7 gene involved in the pathogenesis of NPC.

Preparation of polyclonal antibody specific for BRD7 and detection of its expression pattern in the human fetus.

BRD7 is a novel bromodomain gene. It plays critical role in cell growth, cell cycle progression, and signal-dependent gene expression. Overexpression of the BRD7 gene in nasopharyngeal carcinoma cells is effective to inhibit cell growth and cell cycle progression from G1 to S phase. However, little is known about its bio-functions because of the unavailability of a specific BRD7 antibody. In this study, for the first time, we generated a highly specific BRD7 antibody. It is able to specifically recognize recombinant GST-BRD7N protein with a molecular mass of 65 kDa and recognize BRD7-Myc and endogenously expressed BRD7 protein with an approximate molecular mass of 75 kDa, which corresponds well with the calculated molecular mass of the BRD7 protein. More importantly, with these antisera, we analyzed BRD7 distribution in the human fetus by Western blot and immunohistochemistry assays. Obvious nuclear expression of BRD7 protein presents in human cerebellum, pancreas, intestines, liver, and kidney. Cardiomyocyte shows high cytoplasm expression of the BRD7 protein. Weak nuclear expression of the BRD7 protein is found in human cerebrum, lung, and stomach. These data may help to further study the cellular role of the BRD7 gene. In particular, the prepared BRD7 antibody will be helpful for studying the bio-functions of endogenously expressed BRD7 protein.

BRD2 is one of BRD7-interacting proteins and its over-expression could initiate apoptosis.

BRD7 is a potential nuclear transcription regulation factor related to nasopharyngeal carcinoma (NPC). BRD2, a putative BRD7-interacting protein, has been screened from human fetal brain cDNA library by yeast two-hybrid system. This study was to further identify the interaction between BRD7 and BRD2 in mammalian cells, and to investigate the subcellular localization of BRD2, as well as the effect on the functions of cell biology. Both immunoprecipitation and subcellular colocalization were performed together to identify the interaction of BRD7 with full-length BRD2, as well as C-terminal truncated BRD2 or N-terminal truncated BRD2. GFP direct fluorescence and Hochest 33258 staining were used to investigate the cellular localization pattern of BRD2 and the roles in initiating cell apoptosis in COS7 and HNE1. The results showed that BRD7 could interact with BRD2 and the region from amino acid 430 to 798 of BRD2 was critical for the interaction of BRD2 with BRD7. BRD2 mainly localizes in nucleus in two distribution patterns, diffused and dotted, and BRD2 has distinct roles in initiating apoptosis, and the dotted distribution pattern of BRD2 in nucleus may be a morphologic marker of cell apoptosis.